Domain Architecture Evolution of Metazoan Proteins

Abstract

Appearance of novel multidomain proteins with novel domain architectures (DAs) is closely associated with biological innovations therefore there is a growing interest in genome-scale analysis of the evolutionary history of multidomain proteins. A prerequisite of these studies, however, is that the protein sequences compared are correct and that their evolutionary relationship is correctly defined. We have shown that in the case of most Metazoan proteomes the contribution of mispredicted sequences to apparent DA differences of orthologous and paralogous proteins is greater than the contribution of true gene rearrangements and that analyses of DA evolution frequently suffer from confusing paralogous multidomain proteins (that evolved via gene duplication) with epaktologous multidomain proteins (that are related only through the independent acquisition of the same domain types). Since these methodological errors preferentially increase apparent terminal DA change they may lead to the erroneous conclusion that gene fusion played a dominant role in DA evolution of Metazoan proteins, whereas the contribution of exon shuffling was negligible. In contrast with such conclusions, our studies on high quality datasets of orthologous and paralogous proteins have confirmed that exon shuffling played a major role in the evolution of multidomain proteins of Metazoa.